AGRN: Accurate Gene Regulatory Network inference using ensemble machine learning methods

Table of Content

• Getting Started
• Dataset
• Prerequisites
• Download and install code
• Demo

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.

Dataset

The dataset can be found in the dataset directory. Both DREAM4 and DREAM5 datasets are inside the AGRN_tool.zip tool

the data which have been used in this paper is from DREAM4 and DREAM5 challenges.

In DREAM4 challenge, the dataset comprises two main files:

1-  Gene expression data
2-  GoldStandardard for each network

Network	Organism	# Genes	# Chips
Network 1	In silico	100	100
Network 2	In silico	100	100
Network 3	In silico	100	100
Network 4	In silico	100	100
Network 5	In silico	100	100

In DREAM5 challenge, the dataset comprises three main files:

1-  Gene expression data
2-  GoldStandardard for each network
3-  List of  transcription factors (TFs)

Network	Organism	# TranscFactors	# Genes	# Chips
Network 1	In silico	195	1643	805
Network 2	S.aureus	99	2810	160
Network 3	E. coli	334	4511	805
Network 4	S. cerevisiae	333	5950	536

Prerequisites

You would need to install the following software before replicating this framework in your local or server machine.

Python version 3.7.4

Poetry version 1.1.12

You can install poetry by running the following command:

curl -sSL https://raw.githubusercontent.com/python-poetry/poetry/master/get-poetry.py | python -

To configure your current shell run `source $HOME/.poetry/env`

Download and install code

• Retrieve the code

git clone https://github.com/DuaaAlawad/AGRN.git

Demo

To run the program, first, set the input path in the input.txt file. Here is a sample input file form DREAM5 E-coli Dataset.

./Dataset/DREAM5/training data/Network 1 - in silico/net1_expression_data.tsv
./Dataset/DREAM5/test data/DREAM5_NetworkInference_GoldStandard_Network1 - in silico.tsv
./Dataset/DREAM5/training data/Network 1 - in silico/net1_transcription_factors.tsv
./Dataset/DREAM5/training data/Network 1 - in silico/net1_gene_ids.tsv

Then, run following python command from the root directory.

cd AGRN
poetry install
poetry run python AGRN.py

• Finally, check output folder for results. The output directory contains importance scores from ETR, SVR and RFR in csv files. The OutputResults.txt file shows the results in AUROC and AUPR.

---

To run AGRN using your own data, you should set the input path in the input.txt file.

• the first line in the input file: *expression_data.csv represents the row as samples and the column as genes.
• the second line in the input file: *GoldStandard.csv represents the first column as a transcription factor, the second column as a target gene, and the third column refer to 1 as an interaction pair and 0 as a non-interaction pair.
• the third line in the input file: *transcription_factors.csv contains one column which represents the transcription factors.

Authors

Duaa Mohammad Alawad, Md Wasi Ul Kabir, Ataur Katebi, Md Tamjidul Hoque. For any issue please contact: Md Tamjidul Hoque, [email protected]